Development of a smart variable rate sprayer using deep convolutional neural networks for site-specific application of agrochemicals
Bibliographic record
Abstract
Potato production in Canada typically involves approximately 20 uniform applications (UA) of agrochemicals during a growing season, while usually ignoring spatial and temporal variations in the occurrence of weeds and diseased plants within potato fields. However, UA poses a serious threat to the environment and substantially increases the cost of crop production. Spatial distribution of weeds and diseased plant patches within potato fields emphasizes the need to develop a smart variable rate sprayer (SVRS). Innovations in development of precision agriculture technologies have enabled Engineers\nto develop SVRS using machine vision (MV) and deep learning (DL) to accurately identify and encounter the targets (weeds and diseased plants) in real-time for within-fields variable rate application (VA) of herbicides and fungicides. Five potato fields were selected to collect images of spatially and temporally variable healthy potato plants, diseased potato plants, weeds and their combinations among them and with bare soil patches. The images were collected using a Canon PowerShot SX540 HS camera and Logitech C270 HD Webcam under varying natural light conditions and shadow effects. An image database was constructed by resizing, labeling, processing, and categorizing the above-mentioned images for real-time identification of weed, diseased and healthy plants using DL algorithms. Results of DL models showed > 80% accuracy in detecting targets. The tiny-YOLOv3 models were deployed and integrated into hardware to develop an innovative SVRS (cameras, nozzles, flowmeters, computer, valves and control system). Operational components of the sprayer were calibrated prior to testing in lab and potato fields. The results of lab and field testing revealed that the SVRS was accurate in detecting weeds and diseased plants in real-time and applied agrochemicals on an as-needed basis. Experiments were designed under two-factor factorial arrangements with two treatments (UA and VA) and three levels of weather conditions (cloudy, partly cloudy, and sunny) in a 2x3 factorial design. The spraying techniques and weather conditions were the two independent variables and/or factors of interest with the spray volume consumption as a response variable. A two-way ANOVA test indicated a non-significant effect of the levels of factors of interest on volume consumption of spraying liquid under different weather conditions (cloudy, partly cloudy and sunny); e.g., the spraying application techniques (VA and UA) for both lab and field evaluations had p-values respectively 0.329 and 0.156 for lab testing and 0.968 and 0.751 for field testing during weeds and diseased plant detection experiments. However, there was a significant effect of spraying application techniques on volume consumption (p-value < 0.01). The SVRS was able to save 47 and 51% of agrochemicals for weeds and diseased plant detection experiments, respectively, under all weather conditions. Results indicated that the SVRS was capable of significantly reducing the use of agrochemicals, when compared with UA, both in lab and field environments. The results of this study suggested that the developed SVRS has a great potential to reduce the use of agrichemicals, lower environmental risks, and ultimately improve farm profitability of potato producers.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".